For a DNA molecule, you can define directionality using the convention of a 5'-3' direction on the Sense strand. However, using an N to C convention for a protein doesn't work since the ends may be on the same side and if the protein is unstructured then the N to C direction will change. I'm asking the question from the perspective of addressing PPIs. If I have two proteins that interact with each other, how do you quantitatively describe the relationship of the two proteins with respect to one another. We have formal ways to describing this using domains and residues but not for quaternary structures.
How would you mathematically define "up" for a protein structure?
I would imagine that using RDCs would be able to make some progress into this problem. At the moment it doesn't seem as if there is any convention in the structural biophysics community.